Inference for proportional hazard model with propensity score

Bo Lu, Dingjiao Cai, Luheng Wang, Xingwei Tong, Huiyun Xiang

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Since the publication of the seminal paper by Cox (1972), proportional hazard model has become very popular in regression analysis for right censored data. In observational studies, treatment assignment may depend on observed covariates. If these confounding variables are not accounted for properly, the inference based on the Cox proportional hazard model may perform poorly. As shown in Rosenbaum and Rubin (1983), under the strongly ignorable treatment assignment assumption, conditioning on the propensity score yields valid causal effect estimates. Therefore we incorporate the propensity score into the Cox model for causal inference with survival data. We derive the asymptotic property of the maximum partial likelihood estimator when the model is correctly specified. Simulation results show that our method performs quite well for observational data. The approach is applied to a real dataset on the time of readmission of trauma patients. We also derive the asymptotic property of the maximum partial likelihood estimator with a robust variance estimator, when the model is incorrectly specified.

Original languageEnglish (US)
Pages (from-to)2908-2918
Number of pages11
JournalCommunications in Statistics - Theory and Methods
Volume47
Issue number12
DOIs
StatePublished - Jun 18 2018

Fingerprint

Propensity Score
Partial Likelihood
Cox Proportional Hazards Model
Proportional Hazards Model
Asymptotic Properties
Maximum Likelihood
Assignment
Estimator
Causal Inference
Causal Effect
Cox Model
Observational Study
Right-censored Data
Robust Estimators
Variance Estimator
Confounding
Survival Data
Regression Analysis
Conditioning
Covariates

All Science Journal Classification (ASJC) codes

  • Statistics and Probability

Cite this

Lu, Bo ; Cai, Dingjiao ; Wang, Luheng ; Tong, Xingwei ; Xiang, Huiyun. / Inference for proportional hazard model with propensity score. In: Communications in Statistics - Theory and Methods. 2018 ; Vol. 47, No. 12. pp. 2908-2918.
@article{c83cfaabd6b8467ea7f0cf8943765494,
title = "Inference for proportional hazard model with propensity score",
abstract = "Since the publication of the seminal paper by Cox (1972), proportional hazard model has become very popular in regression analysis for right censored data. In observational studies, treatment assignment may depend on observed covariates. If these confounding variables are not accounted for properly, the inference based on the Cox proportional hazard model may perform poorly. As shown in Rosenbaum and Rubin (1983), under the strongly ignorable treatment assignment assumption, conditioning on the propensity score yields valid causal effect estimates. Therefore we incorporate the propensity score into the Cox model for causal inference with survival data. We derive the asymptotic property of the maximum partial likelihood estimator when the model is correctly specified. Simulation results show that our method performs quite well for observational data. The approach is applied to a real dataset on the time of readmission of trauma patients. We also derive the asymptotic property of the maximum partial likelihood estimator with a robust variance estimator, when the model is incorrectly specified.",
author = "Bo Lu and Dingjiao Cai and Luheng Wang and Xingwei Tong and Huiyun Xiang",
year = "2018",
month = "6",
day = "18",
doi = "10.1080/03610926.2017.1343849",
language = "English (US)",
volume = "47",
pages = "2908--2918",
journal = "Communications in Statistics - Theory and Methods",
issn = "0361-0926",
publisher = "Taylor and Francis Ltd.",
number = "12",

}

Inference for proportional hazard model with propensity score. / Lu, Bo; Cai, Dingjiao; Wang, Luheng; Tong, Xingwei; Xiang, Huiyun.

In: Communications in Statistics - Theory and Methods, Vol. 47, No. 12, 18.06.2018, p. 2908-2918.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Inference for proportional hazard model with propensity score

AU - Lu, Bo

AU - Cai, Dingjiao

AU - Wang, Luheng

AU - Tong, Xingwei

AU - Xiang, Huiyun

PY - 2018/6/18

Y1 - 2018/6/18

N2 - Since the publication of the seminal paper by Cox (1972), proportional hazard model has become very popular in regression analysis for right censored data. In observational studies, treatment assignment may depend on observed covariates. If these confounding variables are not accounted for properly, the inference based on the Cox proportional hazard model may perform poorly. As shown in Rosenbaum and Rubin (1983), under the strongly ignorable treatment assignment assumption, conditioning on the propensity score yields valid causal effect estimates. Therefore we incorporate the propensity score into the Cox model for causal inference with survival data. We derive the asymptotic property of the maximum partial likelihood estimator when the model is correctly specified. Simulation results show that our method performs quite well for observational data. The approach is applied to a real dataset on the time of readmission of trauma patients. We also derive the asymptotic property of the maximum partial likelihood estimator with a robust variance estimator, when the model is incorrectly specified.

AB - Since the publication of the seminal paper by Cox (1972), proportional hazard model has become very popular in regression analysis for right censored data. In observational studies, treatment assignment may depend on observed covariates. If these confounding variables are not accounted for properly, the inference based on the Cox proportional hazard model may perform poorly. As shown in Rosenbaum and Rubin (1983), under the strongly ignorable treatment assignment assumption, conditioning on the propensity score yields valid causal effect estimates. Therefore we incorporate the propensity score into the Cox model for causal inference with survival data. We derive the asymptotic property of the maximum partial likelihood estimator when the model is correctly specified. Simulation results show that our method performs quite well for observational data. The approach is applied to a real dataset on the time of readmission of trauma patients. We also derive the asymptotic property of the maximum partial likelihood estimator with a robust variance estimator, when the model is incorrectly specified.

UR - http://www.scopus.com/inward/record.url?scp=85043356278&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85043356278&partnerID=8YFLogxK

U2 - 10.1080/03610926.2017.1343849

DO - 10.1080/03610926.2017.1343849

M3 - Article

VL - 47

SP - 2908

EP - 2918

JO - Communications in Statistics - Theory and Methods

JF - Communications in Statistics - Theory and Methods

SN - 0361-0926

IS - 12

ER -